There has been an expansion of new medication options for the treatment of patients with bipolar disorder in the last five years. While monotherapy studies are done to win FDA approval of a new drug, few combination studies have been completed, particularly extending beyond a three-week acute period. Despite the lack of evidence combination therapy has become usual clinical practice, as reflected in multiple surveys and publications. We propose a randomized, placebo-controlled study of divalproex (DVP) plus blinded placebo (PBO), lithium carbonate (Li), or quetiapine (QTP) in 105 patients currently hypomanic over a 12-week acute period. We hypothesize the combination treatments (DVP+Li and DVP+QTP) will be associated with greater reduction in symptoms of hypomania and mania than monotherapy alone (DVP+PBO) durng the acute phase 12-week trial. All patients entering will meet criteria for bipolar I disorder and be experiencing at least moderate hypomania. Patients meeting criteria for recovery at 12 weeks will be offered a 14-week continuation on DVP plus blinded medication (PBO, Li, or QTP). This phase will provide exploratory data on the characteristics and course of responders to the initial treatment combinations. Many procedures are incorporated into the study to ensure patient protection. A blinded psychiatrist and research assistant will see patients at weekly intervals for four weeks, and then every two weeks for the duration of their involvement. An unblinded psychiatrist will monitor symptoms, laboratory results, and serum levels. Symptom measures will include the Young Mania Rating Scale, Hamilton Rating Scale for Depression, Clinical Global Assessment for Bipolar Disorder, and other measures of quality of life, functioning, and physical symptoms and side effects. This study will provide fundamental information on the most appropriate initial treatment strategy and preliminary evidence on continuation treatment. We plan a random regression analysis of the primary aim of change of presenting symptoms using a test of slopes (response rates) among the three groups. Secondary aims focus on depressive symptoms and side effects. An exploratory analysis using the Moderator and Mediator approach will seek information on whom these treatments will be most helpful for and what are critical elements within the study design to enhance effect size in future studies on treatment of bipolar disorder.